Abstract-- We present data on the vulnerability of a variety of candidate spacecraft electronics to proton and heavy ion induced single event effects and proton-induced damage. Devices tested include optoelectronics, digital, analog, linear bipolar, hybrid devices, Analog-to-Digital Converters (ADCs), Digital-to-Analog Converters (DACs), and DC-DC converters, among others. I. INTRODUCTION As spacecraft designers use increasing numbers of commercial and emerging technology devices to meet stringent performance, economic and schedule requirements, ground-based testing of such devices for susceptibility to single event effects (SEE) and proton-induced damage has assumed ever greater importance. The studies discussed here were undertaken to establish the sensitivities of candidate spacecraft electronics to heavy ion and proton-induced single event upsets (SEU), single event latchup (SEL), single event transient (SET), and proton damage (ionizing and non-ionizing). The work presented is sponsored by NASA Electronic Parts and Packaging (NEPP) Program's Electronics Radiation Characterization (ERC) Project, Defense Threat Reduction Agency (DTRA) under IACRO 01- 4050/0001278, and NASA Flight Projects. M. V. O'Bryan is with Raytheon Information Technology & Scientific Services, Lanham, MD 20706-4392 (telephone: 301-286-1412, e-mail: [email protected]). Kenneth A. LaBel is with NASA/GSFC, Code 562, Greenbelt, MD 20771 USA (telephone: 301-286-9936, e-mail: [email protected]). Robert A Reed is with NASA/GSFC, Code 562, Greenbelt, MD 20771 USA, (telephone: 301-286-2153, e-mail: [email protected]). II. TEST TECHNIQUES AND SETUP A. Test Facilities All SEE and proton-induced damage tests were performed between February 2000 and February 2001. Heavy Ion experiments were conducted at the Brookhaven National Laboratories (BNL) Single Event Upset Test Facility (SEUTF) and at Texas A&M University Cyclotron (TAMU). The SEUTF uses a twin Tandem Van De Graaf accelerator while the TAMU facility uses an 88" Cyclotron. Both facilities are suitable for providing various ions and energies for testing. At both facilities, test boards containing the device under test (DUT) were mounted in the test area. For heavy ions, the DUT was irradiated with ions with linear energy transfers (LETs) ranging from 0.59 to 120 MeV•cm 2 /mg, with fluences from 1x10 5 to 1x10 7 particles/cm 2 . Fluxes ranged from 1x10 2 to 1x10 5 particles/cm 2 per second, depending on the device sensitivity. Representative ions used are listed in Table I. LETs between the values listed were obtained by changing the angle of incidence of the ion beam on the DUT, thus changing the path length of the ion through the DUT. Energies and LETs available varied slightly from one test date to another. Proton SEE and damage tests were performed at three facilities: the University of California Davis (UCD) Crocker Nuclear Laboratory (CNL), Tri-University Meson Facility (TRIUMF), and the Indiana University Cyclotron Facility (IUCF). Proton test energies incident on the DUT are listed in Table II. Typically, the DUT was irradiated to a fluence from 1x10 10 to 1x10 11 particles/cm 2 , with fluxes on the order of 1x10 8 particles/cm 2 per second. Recent Radiation Damage and Single Event Effect Results for Candidate Spacecraft Electronics Martha V. O’Bryan 1 , Member, IEEE, Kenneth A. LaBel 2 , Member, IEEE, Robert A. Reed 2 , Member, IEEE, Ray L. Ladbury 3 , Member, IEEE, James W. Howard Jr. 4 , Senior Member, IEEE, Stephen P. Buchner 5 , Member, IEEE, Janet L. Barth 2 , Senior Member, IEEE, Scott D. Kniffin 3 , Member, IEEE, Christina M. Seidleck 1 , Cheryl J. Marshall 2 , Member, IEEE, Paul W. Marshall 5 , Member, IEEE, Hak S. Kim 4 , Donald K. Hawkins 2 , Martin A. Carts 1 , James D. Forney 4 , Anthony B. Sanders 2 , Stephen R. Cox 2 , Curtis J. Dunsmore 4 , and Christopher Palor 3 1. Raytheon Information Technology & Scientific Services, Lanham, MD 20706-4392 2. NASA/GSFC, Code 562, Greenbelt, MD 20771 3. Orbital Sciences Corporation, McLean, VA 4. Jackson & Tull Chartered Engineers, Washington, D. C. 20018 5. NRL/SFA Inc, Largo, MD 20785 6. Consultant
Transcript
Abstract-- We present data on the vulnerability of a
variety of candidate spacecraft electronics to proton andheavy ion induced single event effects and proton-induceddamage. Devices tested include optoelectronics, digital,analog, linear bipolar, hybrid devices, Analog-to-DigitalConverters (ADCs), Digital-to-Analog Converters(DACs), and DC-DC converters, among others.
I. INTRODUCTION
As spacecraft designers use increasing numbers ofcommercial and emerging technology devices to meetstringent performance, economic and schedule requirements,ground-based testing of such devices for susceptibility tosingle event effects (SEE) and proton-induced damage hasassumed ever greater importance.
The studies discussed here were undertaken to establish thesensitivities of candidate spacecraft electronics to heavy ionand proton-induced single event upsets (SEU), single eventlatchup (SEL), single event transient (SET), and protondamage (ionizing and non-ionizing).
The work presented is sponsored by NASA Electronic Parts and
Packaging (NEPP) Program's Electronics Radiation Characterization (ERC)Project, Defense Threat Reduction Agency (DTRA) under IACRO 01-4050/0001278, and NASA Flight Projects.
M. V. O'Bryan is with Raytheon Information Technology & ScientificServices, Lanham, MD 20706-4392 (telephone: 301-286-1412, e-mail:[email protected]).
Kenneth A. LaBel is with NASA/GSFC, Code 562, Greenbelt, MD 20771USA (telephone: 301-286-9936, e-mail: [email protected]).
Robert A Reed is with NASA/GSFC, Code 562, Greenbelt, MD 20771USA, (telephone: 301-286-2153, e-mail: [email protected]).
II. TEST TECHNIQUES AND SETUP
A. Test FacilitiesAll SEE and proton-induced damage tests were performed
between February 2000 and February 2001. Heavy Ionexperiments were conducted at the Brookhaven NationalLaboratories (BNL) Single Event Upset Test Facility(SEUTF) and at Texas A&M University Cyclotron (TAMU).The SEUTF uses a twin Tandem Van De Graaf acceleratorwhile the TAMU facility uses an 88" Cyclotron. Bothfacilities are suitable for providing various ions and energiesfor testing. At both facilities, test boards containing the deviceunder test (DUT) were mounted in the test area. For heavyions, the DUT was irradiated with ions with linear energytransfers (LETs) ranging from 0.59 to 120 MeV•cm2/mg, withfluences from 1x105 to 1x107 particles/cm2. Fluxes rangedfrom 1x102 to 1x105 particles/cm2 per second, depending onthe device sensitivity. Representative ions used are listed inTable I. LETs between the values listed were obtained bychanging the angle of incidence of the ion beam on the DUT,thus changing the path length of the ion through the DUT.Energies and LETs available varied slightly from one testdate to another.
Proton SEE and damage tests were performed at threefacilities: the University of California Davis (UCD) CrockerNuclear Laboratory (CNL), Tri-University Meson Facility(TRIUMF), and the Indiana University Cyclotron Facility(IUCF). Proton test energies incident on the DUT are listedin Table II. Typically, the DUT was irradiated to a fluencefrom 1x1010 to 1x1011 particles/cm2, with fluxes on the orderof 1x108 particles/cm2 per second.
Recent Radiation Damage and Single EventEffect Results for Candidate Spacecraft
Electronics
Martha V. O’Bryan1, Member, IEEE, Kenneth A. LaBel2, Member, IEEE,Robert A. Reed2, Member, IEEE, Ray L. Ladbury3, Member, IEEE,
James W. Howard Jr.4, Senior Member, IEEE, Stephen P. Buchner5, Member, IEEE,Janet L. Barth2, Senior Member, IEEE, Scott D. Kniffin3, Member, IEEE, Christina M. Seidleck1,
Cheryl J. Marshall2, Member, IEEE, Paul W. Marshall 5, Member, IEEE,Hak S. Kim4, Donald K. Hawkins2, Martin A. Carts1, James D. Forney4, Anthony B. Sanders2,
Stephen R. Cox2, Curtis J. Dunsmore4, and Christopher Palor3
3. Orbital Sciences Corporation, McLean, VA4. Jackson & Tull Chartered Engineers, Washington, D. C. 20018
5. NRL/SFA Inc, Largo, MD 207856. Consultant
The pulsed laser facility at the Naval Research Laboratorywas used to generate single event transients in integratedcircuits. The laser light used had a wavelength of 590 nm thatresulted in a skin depth (depth at which the light intensitydecreased to 1/e - or about 37% - of its intensity at thesurface) of 2 microns.
Table I: Heavy Ion Test FacilitiesFacility
Brookhaven National Laboratories (BNL)Single Event Upset Test Facility (SEUTF)Texas A&M University Cyclotron (TAMU)
Table IV: Other Test FacilitiesNaval Research Laboratory (NRL) Pulsed Laser SEE Test FacilityLaser: 590 nm, 3 ps pulse width, beam spot size ~1.5 µmGoddard Space Flight Center Radiation Effects Facility(GSFC REF)
B. Test MethodUnless otherwise noted, all tests were performed at room
temperature and with nominal power supply voltages.
1) SEE Testing - Heavy IonDepending on the DUT and the test objectives, one or
more of three SEE test methods were used:Dynamic – the DUT was exercised continually while being
exposed to the beam. The errors were counted, generally bycomparing DUT output to an unirradiated reference device orother expected output. In some cases, the effects of clockspeed or device modes were investigated. Results of such
tests should be applied with caution because device modesand clock speed can affect SEE results.
Static – the DUT was loaded prior to irradiation; data wereretrieved and errors were counted after irradiation.
Biased (SEL only) – the DUT was biased and clockedwhile ICC (power consumption) was monitored for SEL orother destructive effects. In some SEL tests, functionality wasalso monitored.
In SEE experiments, DUTs were monitored for soft errors,such as SEUs and for hard errors, such as SEL. Detaileddescriptions of the types of errors observed are noted in theindividual test results.
SET testing was performed using a high-speedoscilloscope. Individual criteria for SETs are specific to thedevice being tested. Please see the individual test reports fordetails. [1]
Heavy ion SEE sensitivity experiments includemeasurement of the saturation cross sections and the LinearEnergy Transfer (LETth) threshold (the minimum LET valuenecessary to cause an effect at a fluence of1x107 particles/cm2).
2) SEE Testing - ProtonProton SEE tests were performed in a manner similar to
heavy ion exposures in many regards. Differences includemeasuring the SEE cross section as a function of protonenergy as opposed to LET, as well as differences incumulative fluence and particle flux rates.
3) Proton Damage TestingProton damage tests were performed on biased devices
with functionality and parametrics being measured eithercontinually during irradiation or after step irradiations (forexample, every 10 krad (Si), or every 1x1010 protons).
A proton test "lessons learned" document is currently underdevelopment. [2] Optocoupler characterization approachesused in this study are found in [1]
4) Pulsed Laser Facility TestingThe laser light used had a wavelength of 590 nm that
resulted in a skin depth (depth at which the light intensitydecreased to 1/e - or about 37% - of its surface intensity) of2 microns. A pulse rate of 100 Hz was chosen. The DUT wasmounted on an X-Y stage in front of a 100x lens thatproduced a spot size of about 1.5 microns. The X-Y stagecould be moved in steps of 0.1 micron for accuratepositioning of SEU sensitive regions in front of the focusedbeam. An illuminator together with a CCD camera andmonitor were used to image the area of interest, therebyfacilitating accurate positioning of the device in the beam.The pulse energy was varied with a neutral density filter andthe energy was monitored by splitting off a portion of thebeam and directing it at a calibrated energy meter.
5) TID TestingTID testing was performed using a Co-60 source at the
Goddard Space Flight Center Radiation Effects Facility(GSFC REF). The source is capable of delivering a dose rate
of 0.5 Rad(Si)/s, with dosimetry being performed by an ionchamber probe.
III. TEST RESULTS OVERVIEW
Abbreviations and conventions are listed in Table III.Abbreviations for principal investigators (PIs) are listed inTable IV. SEE test results are summarized in Table V.Unless otherwise noted, all LETths are in (MeV•cm2/mg) andall cross sections are in cm2/device. This paper is a summaryof results. Complete test reports are available online athttp://radhome.gsfc.nasa.gov [1].
TABLE V: ABBREVIATIONS AND CONVENTIONS:H = heavy ion testP = proton test (SEE)LET = linear energy transfer (MeV•cm2/mg)LETth = linear energy transfer threshold (the minimum LET value for
which a given effect is observed for a fluence of 1x107
particles/cm2 – in MeV•cm2/mg)LETmax = highest tested LETLETeff = effective LETSEU = single event upsetSEL = single event latchupSET = single event transientSEFI = single event functional interruptDD = displacement damage< = SEE observed at lowest tested LET> = No SEE observed at highest tested LETTID = total ionizing doseσ = cross section (cm2/device, unless specified as cm2/bit)σSAT = saturation cross section at LETmax (cm2/device, unless specified
as cm2/bit)LDC = lot date codeCTR = current transfer ratioDAC = digital to analog converterADC = analog to digital converterLED = light emitting diodeDRAM = dynamic random access memorySRAM = static random access memoryMOSFET = metal oxide semiconductor field effect transistorPAL = Programmable array logic deviceALU = Arithmetic Logic UnitVS = Supply voltageIF = Forward currentVF = Forward voltageVIN = Input voltageN/A = Not applicablep-p = peak-to-peakCat = category
TABLE VI: LIST OF PRINCIPAL INVESTIGATORS
Principal Investigator (PI) AbbreviationKenneth LaBel KLRobert Reed RRJim Howard JHRay Ladbury RLScott Kniffin SKTony Sanders TS
TABLE VII: LIST OF CATEGORIES
Category Implications1 Relatively hard or immune to SEEs;
recommended for spaceflight2 Somewhat susceptible to SEEs; may need
some error detection and correction (EDAC)when used in an application
3 Fairly soft devices that are very susceptible toSEEs; use with great caution. Intensive EDACmay be necessary
4 Not recommended for spaceflight. Destructiveconditions, such as latchup, total dose failure, orburnout were observed in these devices at lowlevels.
TABLE VII: SUMMARY OF SEE TEST RESULTS
Part Number Function LDC Manufacturer Particle:(Facility)P.I.
TestingPreformed
Summary of Results Cat
ADCs:AD1674 ADC 9848 Analog Devices H: (BNL) JH/RR SEU; SEL SEL LETth >37, s sat<1x10-7. For
positive input: SEU LETth<2.6s sat<1.6x10-3. For negative input:SEU LETth=5.2 s sat=2.2x10-4
2
AD6640 ADC 9951 Analog Devices H:(BNL) JH/RL/RR SEU; SEL SEUs LETth < 1; s ~ 1-2x10-3;
s ~ 1-2x10-4cm2 per bit;SEFIs observed; 26.2 < LETth <37;s ~ 1x10-5;No SELs observed; SEL LETth >37
2
SPT7760 ADC 1GSPS
N/A SignalProcessingTech.
H:(TAMU)RR SEU; SEFIFunctionalFailure
Functional failure occurred atLET=84, power cycle was requiredto recover. s=2x10-2 for functionalfailure. No functional failure atLET=50 to a fluence of 1x107 p/cm2
LETth<1.8, s sat~2x10-3cm2/device
2
DC-DC Converters:
AFL12005SX/CH DC-DC 0006 Lambda H: (BNL) JH/RR SET For Cl(LET=11.4) No transientsobserved. Saw destructive failureunder low-voltage, high-loadconditions during irradiation ofsector 2. Sectors 1,3,4 not tested.
3
AFL12012DX/CH DC-DC 0005 Lambda H: (BNL) JH/RR SET For Br(LET=37) No transients ordestructive events observed for highvoltage low load. Saw destructivefailure under low-voltage, high-loadconditions during irradiation ofsector 1.
3
AFL12015DX/CH DC-DC 0005 Lambda H: (BNL) JH/RR SET For Br(LET=37) No transientsobserved. No failure in sectors1,3,4. Failure at low-voltage high-load in sector 2 led to a destructiveevent.
3
DVHF2803R3SF DC/DC 9910 Virginia PowerTechnology
H:(BNL) JH SET/DD For Br(LET=37) No Destructiveevents observed. Destructive eventat LET=60 and full loading. SmallSETs observed at low rate.
SETs 50-100 µs in duratiionP: No SETs observed to s <3x10-12
P: (UCD) TID failure observed atproton doses of ~ 20 and 25 krads (Si)
2
MDI3051RES05ZF DC-DC 0013 Modular
Devices Inc.
H:(BNL) JH/RR SET For Cl(LET=11.4) No transientsobserved. Sector 2: failure at low-voltage high-load led to adestructive event.
3
MDI3051RES12ZF DC-DC 0013 Modular
Devices Inc.
H:(BNL) JH/RR SET For Cl(LET=11.4) No transients orfailures. For Ni(LET=28) no SETs inSector 2: failure at 120 V high-loadled to a destructive event.
3
MDI3051RES15ZF DC-DC 0013 Modular
Devices Inc.
H:(BNL) JH/RR SET For Cl(LET=11.4) No transients orfailures. For Ni(LET=28) no SETsobserved. Sector 2: failure at 120 Vmoderate-load led to a destructiveevent.
3
TABLE VII (CONT.): SUMMARY OF SEE TEST RESULTS
Part Number Function LDC Manufacturer Particle:(Facility)P.I.
TestingPreformed
Summary of Results Cat
Linear Bipolar:
LM-124 Op Amp N/A NationalSemiconductor
Laser:(NRL) JH SET Several SETs observed, see poster[3] for application specific testresults
2
LM-139 Comparator N/A NationalSemiconductor
H: (BNL) JH SET Transients observed; depended oninput differential voltage, see poster[3]
2
HS-139 Comparator N/A Harris/Intersil H:(BNL) JH SET Transients observed; depended oninput differential voltage, see poster[3]
2
LM-148 Op Amp 9905 Fairchild P:(UCD) SK/RR SET;SEU HI: No SEL to I(LET=60); SETLETth<2.6, s sat=2x10-5; Variouspulse height and widthP: No SETs observed, no SEUsobserved; no performancedegradation to 7.44x1011
protons/cm2
2
Board Tests:
Pentium III Processor N/A Intel P:(IU) JH
H: (TAMU) JH
SEE; TID; SEL: No SELs observed to LET>20;No SELs observed for 200 MeVproton; SEU: non-destructive errorswere observed, P: SEFI: wereobserved. See poster [4]
3
AMD K7 Processor N/A AdvancedMicro Devices
P:(IU) JH
H: (TAMU) JH
SEU/SEL No SELs observed; SEU: non-destructive errors were observed,P: SEFI: were observed. See poster[4]
4
Miscellaneous:
IL710 Isolator 9950 Non VolatileElectronics
H:(BNL) RR SEL; SET No SEL observed LETth>60, SETsobserved beginning near LET=11,s sat=2.5x10-5
2
Mii42142 Power OpAmp
N/A Micropac H:(BNL) SK/RRP:(UCD) SK/RR
SET; SEGR HI:SETth<11.4 (dependent on bias)conditions; LETth for SEGRth<37 (application specific)P: No SETs obseved
4
Mii53124 PowerMOSFETOptocoupler
N/A Micropac H:(BNL) SK/RR SET; SEGR No SET;LETth for SEGR = 37
3
Mii53250 Relay N/A Micropac H:(BNL) SK/RR SET; SEGR No SET;LETth for SEGR >60
3
Mii53253 PowerMOSFETOptocoupler
N/A Micropac H:(BNL) SK/RR SEGR No SET;LETth for SEGR = 60
3
Mii53258 Relay N/A Micropac H:(BNL) SK/RR SET; SEGR No SET;LETth for SEGR = 60
3
6651 Optocoupler 0027 Agilent (HewlettPackard)
P:(UCD) SK SET SETs observed with angulardependence
3
TABLE VIII: SUMMARY OF DISPLACEMENT DAMAGE (DD) TEST RESULTS
Part Number Function LDC Manufacturer Particle:(Facility)P.I.
TestingPreformed
Summary of Results
Displacement Damage Test Results:
OD800 LED N/A Optodiode P:(UCD) SK/RR DD Output power noted 20%degradation at 6.6x1010 p/cm2
Micropac P:(UCD) SK DD CTR degradation observed at7.5x1011 p/cm2 (flight lot); and at>1x1012 (non flight lot)
Mii42142 Power OpAmp
N/A Micropac P:(UCD) SK/RR DD No SET/SEGR to 7.44x1012 p/cm2;18% drop in output current after7.44x1012 p/cm2
Mii53124 PowerMOSFETOptocoupler
N/A Micropac P:(UCD) SK/RR DD No SET, no significant degradationto 1x1012 p/cm2
Mii53250 Relay N/A Micropac P:(UCD) SK/RR DD No SET, no significant degradationto 1x1012 p/cm2
Mii53253 PowerMOSFETOptocoupler
N/A Micropac P:(UCD) SK/RR DD No SET, no significant degradationto 1x1012 p/cm2
Mii53258 Relay N/A Micropac P:(UCD) SK/RR DD No SET, no significant degradationto 1x1012 p/cm2
TABLE IX: SUMMARY OF SEL TEST RESULTS
PartNumber
Function LDC Manufacturer Particle:(Facility)P.I.
TestingPreformed
Summary of Results** *Cat
Linear Bipolar Devices:
CMP402 Comparator 0010 Analog Devices H:(BNL) RL/RR SEL No SEL observed up toLETeff=119.8
1
OP16 Op Amp 9917 Analog Devices H:(BNL) JH/RL SEL No SEL observed up to LETeff=37.3 1
OP37 Op Amp 9925 Analog Devices H:(BNL) JH/RL SEL No SEL observed up to LETeff=59.9 1
OP42 Op Amp 9750 Analog Devices H:(BNL) JH/RL SEL No SEL observed up to LETeff=37.3 1
ADC/DAC:
AD7535 DAC 9804 Analog Devices H:(BNL) JH/RR SEL No SEL to LETeff=74.6 1
AD7564 DAC 9950 Analog Devices H:(BNL) RL/RR SEL No SEL to LETeff=59.9 1
AD7664 ADC 864065-2 Analog Devices H:(BNL) JH SEL SEL LETth ~ 8-10, s = 2-3x10-4 3
AD7854 ADC 9930 Analog Devices H:(BNL) RL/RR SEL SEL 6.7<LETth<11.4, s sat <1x10-3. 4
AD7858 ADC 0026 Analog Devices H:(BNL) RL/RR SEL SEL LETth 11.4-22.8, s sat <1x10-3. 4
AD7888 ADC 0002,
0016
Analog Devices H:(BNL) RL/RR SEL SEL LETth 16.7-22.8, s sat <1.3x10-5. 4
Miscellaneous:
ADG704 Multiplexer 0008 Analog Devices H:(BNL) RL/RR SEL SEL LETth~16, s sat<1.1x10-4. 4
AMP01 Inst. Amp 9922 Analog Devices H:(BNL) RL/KL/JH SEL Hard Failure LETth>32.7; failuremore likely at normal incidenceSusceptible to SET [5]
3
INA117 Diff. Amp N/A TI/Burr-Brown H:(BNL) RL SEL No SEL observed up toLETeff=119.8
1
MAX313 Switch 9823 Maxim H:(BNL) JH/RR SEL No SEL observed up to LETeff=59.9 1
MAX4503 CMOS switch N/A Maxim H:(BNL) JH SEL No SEL observed up toLETeff=119.8
2
MAX4528 CMOSanalogswitches
9816 Maxim H:(BNL) RL/RR SEL No SEL observed up to LETeff=59.9 1
MAX4583 CMOSanalogswitches
0007 Maxim H:(BNL) RL/RR SEL No SEL observed up to LETeff=59.9 1
MAX4617 MUX N/A Maxim H:(BNL) RL SEL No SEL observed up toLETeff=119.8
1
MAX584 Vref 9838 Maxim H:(BNL) RR SEL No SEL observed up to LETeff =37.3
1
Note: * Rating applies only for SEL not for other SEE concerns. ** All parts tested to a minimum fluence of 1.0 x 107 particles/cm2
IV. SEE TEST RESULTS AND DISCUSSION
A. ADCs:
1) AD1674The Analog Devices AD1674 analog-to-digital converter
(ADC) was tested for SEU and SEL at BNL using ions witheffective LETs ranging from 2.6 MeV·cm2/mg (Oxygen) toapproximately 56 MeV•cm2/mg (Bromine at 45 degrees). NoSEL events were observed for normal incidence Br up to afluence of 1x107 ions/cm2, or for a fluence of 8x105 ions/cm2
Br ions incident at 45 degrees. To investigate SEUs in theAD1674, the output of the device under test (DUT) wascompared to that of a reference chip located outside the beamand given an identical input to that of the DUT. Device upsetcross section vs. LET was measured for both positive andnegative inputs to the ADCs. For all tests, the lowest twosignificant bits were masked off to avoid counting noise asupsets.
For positive device input, the threshold LET was about 2-3,and the limiting cross section was approximately 1.6x10-3cm2.
For negative input, the threshold LET for upset was found tobe about 5-8, with a limiting cross section of about2.2x10-4cm2, about an order of magnitude lower than forpositive input. This indicates that the mechanisms underlyingupsets under conditions of positive input and negative inputare probably distinct. For additional information refer to theGSFC radiation group website. [1]
The hardware setup for this testing is shown via a blockdiagram pictured in Figure 1. The data taken for the case of 2masked bits is featured in Figure 2. [6]
Figure 1. Block diagram of the experimental setup for testing the AD1674ADC.
Figure 2. Cross section versus effective LET curve for the AD1674.
2) AD6640The Analog Devices AD6640 12-bit ADC was tested for
SEU, SEL, and single event functional interrupt (SEFI). TheAD6640 is a 65 MHz, 12-bit ADC fabricated in AnalogDevices XFCB dielectrically isolated bipolar process.Because of the high speed required in this test (15-40 MHz),noise was a concern, and several steps were taken to ensurethat noise was kept to a minimum. Even so, several leastsignificant bits had to be masked off to avoid triggering onnoise. The number of masked bits varied from 3 at the lowestfrequencies up to 5 at the higher frequencies. SEUs wereobserved even for carbon ions, which have LET=1.44MeV•cm2/mg, indicating that the LET threshold is less than 1MeV•cm2/mg. The limiting cross section for SEU wasobserved to be about 1-2x10-3cm2 per device. When the cross
sections were calculated per bit, based on the number of bitsavailable to cause upsets (that is 12 - the number of maskedbits), these per bit cross sections gave a fairly tightdistribution with limiting value on the order of 1-2x10-4cm2
per bit.Single event functional interrupts were also observed.
These events caused every conversion to be in error after theyoccurred until they were reset by cycling power to the device.The threshold LET for these SEFIs was between 26.2 and 37MeV•cm2/mg and the limiting cross section was on the orderof 1x10-5cm2 per device. No evidence of SEL was observedup to a fluence of 2x106 ions/cm2 for Br79. [7]
3) SPT7760The SPT7760 is an Emitter Coupled Logic (ECL) based
1 giga-sample per second (1 Gsps) 8 bit + overflow flashADC manufactured by Signal Processing Technologies,Incorporated. It has an input range of 0 to 2 V and is fullyparallel with 8 bits of resolution (256 values) plus anadditional over-range bit. The analog input has a bandwidthof over 900 MHz and a capacitance of < 15 pF.
The output byte-stream is at differential ECL levels. It isde-multiplexed into two identical ports (labeled ports A andB) each with 9 bits (8 + overflow) plus clock at a maximumoutput rate of 500 Msps. The aggregate throughput istherefore 1 Gsps. The device is supplied in an 80 pinMQUAD package with the option of MIL-STD-883screening.
The SPT7760 was tested for susceptibility to heavy ioninduced SEUs at TAMU using beams of Ne, Kr and Xe ions.Because the input was a constant dc voltage, the output fromthe ADC was also constant unless an SEU occurred. As aresult, SEUs were identified by comparing each conversion tothe previous conversion. Any changes were recorded asSEUs and the data were stored on a PC for post-processing.[8].
Figure 3 shows results from heavy ion SEU tests at TAMU.The plot shows measured cross sections for the device forvarious dc input conditions (-2 to 0 volts), frequencies (10,100 and 1000 MHz) and LETs (1.28, 20 and 50). No cleardependence of cross section on input voltage or frequency isevident in the data.
Functional failure occurred at LET=84 and a power cyclewas required to recover. The cross section for functionalfailure was approximately 2x10-2 cm2. No functional failureswere observed up to LET=50 for a fluence of 1x107 p/cm2.The threshold LET for SEU was less than 1.8 and the limitingcross section was on the order of about 2x10-3cm2.
Figure 3. Heavy Ion induced SEU testing on the SPT7760.
B. DC-DC Converters:
1) AFL12005SX/CHThe AFL12005SX/CH Advanced Analog Lambda, Inc.
DC/DC Converter was tested for susceptibilty to destructivesingle event effects and single event transients at BNL usingCl35 ions with an LET of 12.
The DC/DC converter was tested under bias conditions of126 V with a load of approximately 7.7 W and 113 V with aload of approximately 56.4 W. The DC/DC converter was de-lidded and the active device area divided into four circularregions. With the active section of the converter in the lefthalf of the device, the four regions, numbered 1 through 4.Region #2 contained the power MOSFETs.
For the first of the conditions (high voltage and lowloading), no single event transients were observed at theoutput voltage port and no destructive events were observedfor any of the four regions. However, the device experienceda destructive event when region #2 was exposed under thesecond set of conditions. That event resulted in the outputdropping to zero volts and a loss of functionality that was notrecovered after a power cycle. No transient events wereobserved, as the device’s first event was destructive. Thisindicates that the device is susceptible to destructive errorswith a threshold LET less than or equal to 12 with a crosssection that cannot be determined from the current study. [9]
2) AFL12012DX/CHThe AFL12012DX/CH Advanced Analog Lambda, Inc.
DC/DC Converter was tested for single event destructive andtransient susceptibility at BNL using Br79 ions with an LET of37.
The DC/DC converters were tested under bias conditionsof 120 and 126 V with a load of approximately 7.7 W and116 V with a load of approximately 57.6 W. The DC/DCconverter was de-lidded and the active device area dividedinto four circular regions. With the active section of theconverter in the left half of the device, the four regions,numbered 1 through 4, start in the upper left hand corner andproceed clockwise. The power MOSFETs were primarilylocated in region #2, but may have overlapped into region #1.For the first of the conditions (high voltage and low load), nosingle event transients were observed at the output voltageport and no destructive events were observed for any of the
four regions. However, the device did experience adestructive event when region #1 was exposed under thesecond set of conditions. That event resulted in the outputdropping to zero volts and functionality being lost. Cyclingpower did not return the device to functionality. Thisindicates that this device is susceptible to destructive failurewith a threshold LET less than 37 and a cross section thatcannot be determined from this study. The susceptibility ofthe AFL12005SX/CH) to destructive failure with thresholdLET less than 12 may indicate that the threshold LET issubstantially lower than 37. [10]
3) AFL12015DX/CHThe AFL12015DX /CH Advanced Analog Lambda, Inc.
DC/DC Converter was tested for single event destructive andtransient susceptibility at BNL using ions with an LET of 37(Br79).
The DC/DC converters were tested under bias conditionsof 126 V with a load of approximately 7.8 W and 116, 117,and 113 V with a load of approximately 55.6 to 58.5 W. TheDC/DC converter was de-lidded and the active device areadivided into four circular regions. With the active section ofthe converter in the left half of the device, the four regions,numbered 1 through 4, start in the upper left hand corner andproceed clockwise. Region #2 contained the powerMOSFETs.
For the first of the conditions (high voltage and low load),no single event transients were observed at the output voltageport and no destructive events were observed for any of thefour regions. However, the device did experience adestructive event when region #2 was exposed under biasconditions of 116, 117, and 113 V at a loading ofapproximately 55.6 to 58.5 W. That event resulted in a loss ofdevice functionality and a drop of the device output to zerovolts. Functionality could not be restored by cycling power.No transient events were observed, as the device’s first eventwas destructive. This result indicates that the device issusceptible to a destructive failure with a threshold LET lessthan or equal to 37, and with a cross section that cannot bedetermined from the current study. The fact that another AFLpart (AFL12005SX/CH) exhibited a similar destructivefailure at an LET of 12 suggests that the threshold LET couldbe considerably lower than 37. [11]
4) DVHF2803R3SFThe DVHF2803R3SF/HBM DC/DC Converters from
Virginia Power Technology, Inc. were tested to determinetheir susceptibility to destructive single event effects andsingle event transients at BNL using ions with LETs of 37(Br79) and 59.8 (I127).
The DC/DC converters were tested under bias conditionsof 28 and 35 V with loads of approximately 1.65 and 9.9 W.The DC/DC converter was de-lidded and the active devicearea was divided into three circular regions. Region 1contains the power devices and is located near the upperright-hand corner, region 2 is the upper left-hand side and
region 3 the lower left-hand side. The lower right containedonly passive devices.
For these converters both destructive events and transientswere observed, albeit at high LET values. Because thedestructive events may be of greater concern, they will bediscussed first. DUTs 1 and 2 were test with Bromine ions(LET = 37) at low and high load conditions and with bothnominal (28 V) and high (35 V) input conditions. Under all ofthese conditions, no destructive events were observed. Forruns with Iodine ions (LET = 59.8), destructive events wereobserved. However, these events occurred only when thedevice was operated at full load. Both DUTs passed at Iodineunder low load conditions. Additionally, DUT 2 passed onerun with input voltage 28 V and high load with Iodine, butfailed on the next run when the input voltage was raised to 35V. The location of the power devices and the structuressurrounding them made it impossible to irradiate the parts atangles other than normal incidence, so effective LETsbetween the nominal ion LETs could not obtained. Therefore,the threshold LET for destructive events is between 37 and59.8. For the two destructive events observed, the averagecross section for the event is approximately 2 x 10-6cm2,although the low statistics and uncertainties in the finalfluence dictate large error bars for this number.
Transient errors were also observed as positive pulses ofmagnitude up to 500 mV superimposed on the output voltage.Transient durations were on the order of a microsecond,although the exact duration could not be determined becauseof limitations on the timescales of the digitizing scope.Transients were observed on both devices (although DUT 2was approximately an order of magnitude more sensitive) atboth LET = 37 and 59.8 MeV•cm2/mg. This indicates that thethreshold LET for these events is less than 37. No lowerbound can be determined from the current data. However,cross sections were small (10-7 to 10-5 cm2), suggesting afairly low event rate even for low LET thresholds. Moreover,the short duration of the transients suggests that they could besubstantially mitigated by filtering. These considerationslowered the priority of exact determination of the LETthreshold for these transients. [12]
5) LCM-120SET and SEL testing were performed on the Interpoint
LCM-120 line conditioning module. This device was tested atthree facilities, BNL, Naval Research Laboratory PlusedLaser SEE test facility, and University of California CrockerNuclear Laboratory.
Based on these test runs, it has been determined that theLCM-120 is susceptible to single event transients on theoutput of the device when exposed to heavy ions. Thethreshold for the transients is <2.6 (lowest LET used) and thesaturation cross section is approximately 10-3 cm2. Thetransients all lasted on the order of 50-100 µs. The transientstook the form of voltage drops from the nominal outputvoltage of 24 V, and they ranged from a volt to approximately18 V. The large voltage transients exhibited very rapid fall
times, and because of limitations of the digital scope tocapture such rapid events, it cannot be ruled out that the dropwas all the way to ground.
Despite the very low threshold LET observed for heavyions, proton testing at UCD did not reveal any proton inducedSET sensitivity to a cross section of less than 3 x 10-12 cm2.The sensitivity to proton-induced SETs at lower crosssections could not be evaluated due to the total doserestrictions of the LCM-120. Total dose failure was observedat the UCD facility at proton doses of approximately 20 and25 krads (Si) for the two devices tested there [13]. TIDtesting was performed on the LCM-120, at the GSFC-REFusing Co-60 ? rays. The device exceeded specificationsbetween 23-28 krads (Si) [14].
10-6
10-5
10-4
10-3
403020100
LET (MeV-cm2/mg)
Region #2Region #3
Cro
ss S
ecti
on
(cm
2 )
10-6
10-5
10-4
10-3
10-6
10-5
10-4
10-3
403020100 403020100
LET (MeV-cm2/mg)
Region #2Region #3Region #2Region #3
Cro
ss S
ecti
on
(cm
2 )
Figure 4. Plot of LCM-120 SET cross section as a function of the ion LET.Data for two Device s Under Test (DUT) are shown.
Figure 5. Photograph of the LCM-120 device with the four sensitivecomponents circled in yellow.
6) MDI3051RES05ZFThe MDI3051RES05ZF DC/DC Converter from Modular
Devices, Inc. was tested for susceptibility to single eventinduced transients and destructive events induced at the BNL.
The DC/DC converter was tested under bias conditions of126 V with a load of approximately 7 W, and 113 V with aload of approximately 52.9 W. The DC/DC converter was de-lidded and the active device area divided into two circularregions. With the active section of the converter in the upperleft quadrant of the device, region 1 is in the upper left-handcorner and region two is to the immediate right of region 1.Region #2 contained the power MOSFETs.
For the first of the conditions (126 V and a load of 7 W),no single event transients were observed at the output voltageport and no destructive events were observed for either of thetwo regions. However, the device did experience a destructiveevent when region #2 was exposed under the second set ofconditions (113 V, 52.9 W load). That event resulted in theoutput dropping to zero volts and the device losingfunctionality. Functionality could not be restored by cyclingpower. No transient events were observed, as the device’sfirst event was destructive. This destructive conditionoccurred for the only LET ion used (LET = 12). These dataindicate that the device is susceptible to a single eventinduced destructive mechanism with threshold LET less than12 and a cross section that cannot be determined from thecurrent study. [15]
7) MDI3051RES12ZFThe MDI3051RES12ZF DC/DC Converter from Modular
Devices, Inc. was tested for susceptibility to single eventinduced transients destructive events at BNL.
The DC/DC converters were tested with an input voltage of120 V and loads of approximately 7.8, 18.9, 37.1 and 55.5 Wusing a beam of Cl ions (LET=12). Testing was also done at120 V and loads of approximately 8, 19.2, 37.6 and 56.3 Wfor a beam of Ni ions (LET=26.7). The DC/DC converter wasde-lidded and the active device area divided into two circularregions. With the active section of the converter in the upperleft quadrant of the device, region 1 was in the upper left-hand corner and region two was to the immediate right ofregion 1. Region 2 contained the power MOSFETs.
No single event transients were observed while irradiatingregion 2 with either ion for any test conditions. Previoustesting had indicated that region 1 was not sensitive. Nodestructive events were observed for the high load conditions.However, the device did experience a destructive event whenregion #2 was exposed to Ni when the input was 120 V andthe load was 56.3 W. That event resulted in the outputdropping to zero volts and the device losing functionality.Functionality could not be restored by cycling power. Notransient events were observed, as the device’s first event wasdestructive. This indicates that the device is susceptible to adestructive condition with a threshold LET between 12 and28. The cross section for this destructive mechanism cannotbe determined from this study. [16]
8) MDI3051RES15ZFThe MDI3051RES15ZF DC/DC Converter from Modular
Devices, Inc. was tested for susceptibility to single eventinduced transients and destructive events at BNL.
The DC/DC converters were tested under conditions of 120V input and loads of approximately 8.2, 19.4, 38.3 and 56.6W for the Chlorine beam (LET=12). Testing was also donefor 120 V input and loads of approximately 8.2, and 38 Wand 75 V input and 38.7 W for the Nickel beam. The DC/DCconverter was de-lidded and the active device area dividedinto two circular regions. With the active section of theconverter in the upper left quadrant of the device, regions 1
was in the upper left-hand corner and region two was to theimmediate right of region 1. Region 2 contained the powerMOSFETs.
No single event transients were observed at the outputvoltage port for any of the conditions tested under irradiationby either ion. No destructive events were observed for region#2 under conditions of high load. (Other testing hadindicated that region #1 was not sensitive). However, thedevice did experience a destructive event with a Nickel beamincident on region #2 while the input voltage was 120 V andthe load was 38 W. That event resulted in the output droppingto zero volts and the device losing functionality.Functionality could not be restored by cycling power. Notransient events were observed, as the device’s first event wasdestructive. These data indicate that this device is susceptibleto a destructive mechanism with an LET threshold between12 and 28. The cross section cannot be determined from thisstudy. [17]
C. Linear Bipolar Devices:
1) LM124, LM139, and HS139The National Semiconductor LM124 Op Amp was tested
for transient susceptibility at Naval Research LaboratoryPlused Laser SEE test facility.
The National Semiconductor LM139, and theHarris/Intersil HS139 Comparators were tested for heavy ioninduced transient susceptibility at BNL.
Transients were observed. For application specific testresults see C. Poivey, et al., “Development of a Test Methodologyfor Single Event Transients (SETs) in Linear Devices” [3].
2) LM148The Fairchild LM148 (LDC9905) quad op amp was
irradiated with 63MeV protons at UCD. The device wasbiased with VS = 15V and IL = 2mA and a 10kHz, 10V peak-to-peak sine wave as an input. No SETs and no performancedegradation were observed up to a fluence of 7.44x1011p/cm2.
The Fairchild LM148 (LDC9905) quad op amp wasirradiated with heavy ions at BNL to determine the SET/SELcharacteristics of the device. Test conditions were identical tothose at UCD. An SET was defined as an output deviation of±0.5V from the expected value of the output sine wave. TheSET LETth = 3.7MeV•cm2/mg and σsat = 8x10-4cm2. No SELswere observed in testing. [18]
Figure 6. Heavy ion SET cross section data for two LM148 devices.
D. Board Tests:
1) Pentium IIIIntel Pentium III processors were tested for SEE and TID
at TAMU, and at the IUCF. No SELs observed for LETs upto at least 20. No SELs observed for 200 MeV protons. Non-destructive SEU errors were observed in various regions ofthe processors and SEFIs were observed (where the processorhalted). See J.W. Howard et al., “Total Dose and SingleEvent Effects Testing of the Intel Pentium III (P3) and AMDK7 Microprocessors”. [4] and [19]
2) AMD K7AMD K7 processors were tested for SEU and SEL at
TAMU and at the IUCF. No SELs were observed. Non-destructive SEU errors were observed in various regions ofthe processors and SEFIs were observed (where the processorhalted). See J.W. Howard et al., “Total Dose and SingleEvent Effects Testing of the Intel Pentium III (P3) and AMDK7 Microprocessors”. [4] and [19]
E. Miscellaneous:
1) IL710The Non Volatile Electronics (NVE) IL710 digital isolator
was tested at BNL to determine its sensitivity to heavy ioninduced single event effects (SEE). Two samples were tested,one manufactured using 1.2 µm fabrication process and theother manufactured using a 0.6 µm process.
During testing we observed that the SETs were typically5µs in duration. Using an iris to modulate the beam size, wewere able to focus the beam on the entire DUT or on eitherthe output or the input side of the DUT. The SET crosssection obtained when exposing the input side of the DUT isbetween 6x10-7 and 1.2x10-6 for an LET = 37 MeV•cm2/mg.No upsets were observed when exposing the output side to afluence of 1x107 ions/cm2. This clearly indicates that theinput side of the device is more sensitive to SETs. NeitherIL710 type experienced SELs for LETs up to 37. [20]
2) Mii42142The Micropac Mii42142 power op amp was irradiated with
heavy ions at BNL to determine devices susceptibility to SETand other SEE. Irradiation took place under several biasconditions, including combinations of VS = 8V or 15V and
V-IN = 0.1V, 0.6V, or 1.3V, V+IN was set to ground. Thedevice was set to a Gain of 10 and inverted output with a 5Ωload. The SET trigger levels were based on V-IN and set at –1.25V, –7.0V and –15V respectively. The device wasdeemed to have experienced a SET if the device outputdeviated from the expected level by the amount set by thetrigger levels and a single event induced destructive effect ifthe device failed to reset after power cycling. The heavy ionSETTH = 11.4 MeV•cm2/mg. The heavy ion σsat = 2x10-4cm2
for transients at V-IN = 0.1V at both values of VS (Figures 7and 8). An SEGR occurred at LET = 60 MeV•cm2/mg withVS = 8V and V-IN = 0.1V. No proton SETs were observed.[21]
Figure 7. Mii42142 SET Cross Section for the indicatedinput voltages at VS = 15V.
Figure 8: Mii42142 SET Cross Section for the indicatedinput voltages and VS = 8V.
3) Mii53124 and Mii53253The Micropac Mii53124 ±90V/0.8A, and Mii53253 Dual
90V/0.8A power MOSFET optocoupler hybrids wereirradiated with heavy ions at BNL to determine theirsusceptibilities to SET and other SEE. All devices showedsome voltage fluctuations on the input that appeared to beprecursors to SEGR. One Mii53124 device did have onecurrent-based anomaly with the LED on. Cycling power resetthe device. This anomaly occurred under irradiation with Br(LET = 37.3) and was not observed under any otherconditions. All devices experienced SEGR, which caused theoutput voltage to drop significantly. Resetting the device did
not return the device to functionality after SEGR. The onsetof this effect occurred at 70V with LET = 60MeV•cm2/mg forboth device types. The device was deemed to haveexperienced a SET if the device switched on with the LEDoff. No SETs were observed for either device type. [22]
solid-state relay was irradiated with heavy ions at BNL todetermine device susceptibility to SET and SEE. The devicewas deemed to have experienced a SET if the device switchedon with the LED off. This device showed no SET or SEGRto a fluence of 1x107 particles/cm2 with LET =60MeV•cm2/mg at normal incidence and at 50° to the normal.The device was tested with V0 = 40V and I0 = 0.6A. This wasdeemed to be the worst-case condition as this represents thelargest voltage drop from drain-to-source across theMOSFET. This device was also tested with the LED on todetermine if there were any other effects; no anomalies wereobserved. [23]
5) Mii53258The Micropac Mii 53258 120VDC/5A(2A) optocoupler
hybrid solid-state relay was irradiated with heavy ions at BNLto determine its susceptibility to SET and other SEE. Thedevice was deemed to have experienced a SET if the deviceswitched on with the LED off. SEGR occurred in one deviceunder irradiation with iodine ions (LET = 59.9) at normalincidence while the device was biased at 120V (the maximumrated voltage for the part) and I0 = 0.2A. No SETs wereobserved in this device. The other devices showed no SET orSEGR to a fluence of 1x107particles/cm2 with Br-81 (LET =37 MeV•cm2/mg) at normal incidence and 50° to the normalwith V0 = 60V and 120V. Nor were SET or SEGR observedfor these devices under irradiation with I-127(LET = 60 MeV•cm2/mg) at normal incidence and 50° withV0 = 60V with and I0 = 0.2A in all cases. [23]
high-speed optocoupler was irradiated with 63MeV protons atUCD to determine the SET response of the device. Thedevice was biased at VCC = 5V, VF = 5V, and IF wasselectable via eight input resistors. During the testing thedevices showed no significant CTR degradation to a fluenceof 1.36x1013p/cm2.
The devices were tested for SETs with three outputconfigurations: passive filter, active filter and no output filter.Four channels in each part were divided into two channelswith no output filter and one channel each with passive andactive filter. An SET was defined as an output deviation of±0.5V from the expected value of the output sine wave. Thedevice channels with the passive filter showed no transientson the output. The device channel with the active filtershowed only a very few output transients with an angle ofincidence effect and a cross section on the order of 10-10cm2
(Figure 9). The channels with no output filter showed a
significant number of SETs with an angle of incidence effectand a maximum cross section of 5.4x10-7cm2 (Figure 10).[24] and [25]
0
1x10-10
2x10-10
3x10-10
4x10-10
5x10-10
6x10-10
7x10-10
0 20 40 60 80 100
Angle of Incidence (Deg)
SE
T C
ross
Sec
tion
(cm
2)
Channel 1Channel 2
Figure 9. SET Cross Section vs. Angle of Incidence for the HCPL6651 withactive SET filter (non-flight lot).
0
1x10-07
2x10-07
3x10-07
4x10-07
5x10-07
6x10-07
0 20 40 60 80 100
Angle of Incidence (deg)
Cro
ss S
ectio
n (c
m2 )
Channel 1Channel 2
Figure 10. SET Cross Section vs. Angle of Incidence for the HCPL6651 withno SET filter (non-flight lot).
V. DISPLACEMENT DAMAGE TEST RESULTS AND DISCUSSION
1) OD800The OD800 LED from Optodiode was irradiated with
63MeV protons at UCD to determine the radiation induceddegradation in the output of the device. The device is a GaAsdouble heterojunction LED. The mean ratio of post-irradiation power to pre-irradiation power, P/P0, for all biasconditions was approximately 0.8 for a single exposure to6x1010p/cm2. [26]
was irradiated with 63MeV protons at UCD to determine theCTR degradation of the device. The device was biased atVCC = 5V, VF = 5V, and IF was selectable via eight inputresistors. There was some degradation in CTR at1x1010p/cm2 and significant degradation was noted by thetime a fluence of 4x1010p/cm2 was reached. No SETs wereobserved. [24]
3) 66099The Micropac 66099 (LDC0048) optocoupler was
irradiated with 63MeV protons at UCD to determine the CTRdegradation of the device. The device was biased at VCC =5V, VF = 5V, and IF was selectable via eight input resistors.The onset of CTR degradation occurred at 2.5x1011 p/cm2 anddevice performance was unacceptable at 7.5x1011p/cm2. NoSETs were observed. [24]
4) 4N49 and 4N49SThe Micropac 4N49 (LDCs 9803 & 9819) was irradiated
with 63MeV protons at UCD to determine the degradation ofVCE. No significant degradation in VCE was observed to afluence of 5x1011p/cm2. Degradation was observed for theworst-case bias condition of VCC = 34V, VIN = 3.6V at1x1011p/cm2.
Figure 11. Micropac 4N49 (LDCs 9803 & 9819) average VCE results for theworst, nominal and best cases up to 5x1010 p/cm2 [27].
The Micropac 4N49 (LDC 0048) optocoupler wasirradiated with 63MeV protons at UCD to determine the CTRdegradation of the device. The device was biased atVCC = 5V, VF = 5V, and IF was selectable via eight inputresistors. The onset of CTR degradation occurred at 5x1010
p/cm2 and device performance was unacceptable at 1.5x1011
p/cm2. No SETs were observed. [24]
Figure 12. CTR degradation in 4N49S DUT 394 (Flight Lot)
Figure 13. CTR degradation in 4N49 DUT 941 (Non-Flight Lot)
Figure 14. Comparison of 4N49 CTR degradation flight lot to non-flight lotfor IF=12.5mA. CTR0 for DUT 394 = 0.623 (flight 4N49S), CTR0 for DUT865 = 0.586 (non-flight 4n49).
The Micropac 4N49S (LDC 9736) optocoupler wasirradiated with 63MeV protons at UCD to determine the CTRdegradation of the device. The device was biased atVCC = 5V, VF = 5V, and IF was selectable via eight inputresistors. The onset of CTR degradation occurred at1x1012 p/cm2 and device performance was unacceptable at4.5x1012 p/cm2. No SETs were observed. [24]
optocoupler was irradiated with 63MeV protons at UCD todetermine the CTR degradation of the device. The devicewas biased at VCC = 5V, VF = 5V, and IF was selectable viaeight input resistors. The onset of CTR degradation occurredat 1.0x1012 p/cm2 and device performance was unacceptableat 3.0x1012 p/cm2. The SET test function did not operatecorrectly during testing, but SETs were expected. [24]
6) 6N140The Micropac 6N140 (66124) (LDC9724) optocoupler was
irradiated with 63MeV protons at UCD to determine the CTRdegradation of the device. The device was biased atVCC = 5V, VF = 5V, and IF was selectable via eight inputresistors. The onset of CTR degradation occurred at7.5x1011 p/cm2 and device performance was unacceptable at1.0x1012 p/cm2. No SETs were observed.
The Micropac 6N140 (66124) (LDC0048) optocoupler wasirradiated with 63MeV protons at UCD to determine the CTRdegradation of the device. The device was biased atVCC = 5V, VF = 5V, and IF was selectable via eight inputresistors. The onset of CTR degradation occurred at1.0x1012 p/cm2 and device performance was unacceptable at3.0x1012 p/cm2. No SETs were observed. [24]
Fluence (p/cm2)Figure 17. Comparison of 6N140 CTR degradation flight lot to non-flight lotfor IF=5.71mA
7) Mii42142The Micropac Mii42142 power op amp was irradiated with
63MeV protons at UCD to determine device susceptibility to
SET. The part showed no transients in its most sensitiveconfiguration of VS = 8V and V-IN = 0.1V. The device alsoshowed no loss of output current in the most stressfulcondition of VS = 15V and V-IN = 1.3V up to3x1012 protons/cm2 and only an 18.5% loss of output currentafter 7.44x1012 protons/cm2. No SEGRs were observed. [21]
8) Mii53124 and Mii53253The Micropac Mii 53124 ±90V/0.8A, and Mii53253 Dual
90V/0.8A power MOSFET optocoupler hybrids wereirradiated with 63MeV protons at UCD to determine devicesusceptibility to SET. The DUTs showed no evidence ofSET. There was a nominal increase in the forward currentthreshold for the LED but the threshold remained below theminimum specification limit for operating the devices. Therewas also no pronounced change in the output currents of thedevices. Each DUT was tested to a total fluence of1x1012 p/cm2. [23]
9) Mii53250 and Mii53258The Micropac Mii53250 40V/10A(2A) and Mii53258
120VDC/5A(2A) optocoupler hybrid solid-state relays wereirradiated with 63MeV protons at UCD to determine theirsusceptibility to SET. The devices exhibited no SETs. Therewas a nominal increase in the forward current threshold forthe LED, but threshold remained below the minimumspecification limit for operating the device. There was alsono pronounced change in the output current of the device.The devices were tested to a total fluence of 1x1012 p/cm2.[22]
VI. SEL TEST RESULTS AND DISCUSSION
A. Linear Bipolar Devices:
1) CMP402Heavy ion irradiations were conducted on the Analog
Devices CMP402 comparator to ensure that they were notsusceptible to SEL or any other SEE induced hard failure. NoSELs were observed for a fluence of 1x107 Iodine ions/cm2
incident at 60o to the normal (LETeff = 119.8).
2) OP16, OP37, and OP42Heavy ion irradiations were conducted on three Analog
Devices op amps, OP42, OP37, and OP16, to ensure that theywere not susceptible to SEL or any other SEE induced hardfailure. All three parts survived irradiation to a fluence of1x107 ions/cm2 of Br at normal incidence (LET=37).
B. ADC/DAC:
1) AD7535Heavy ion irradiations were conducted at BNL to
determine the SEL sensitivity of the Analog DevicesAD7535. No SELs were observed for a fluence of 1x107 ionsof Br incident at 60 to the normal (LETeff = 74.6).
2) AD7564Heavy ion irradiations were conducted on the Analog
Devices AD7564 comparator to ensure that they were not
susceptible to SEL or any other SEE induced hard failure. NoSELs were observed up to a fluence of 1x107 Iodine ions/cm2
incident at 60o to the normal (LETeff = 119.8).
3) AD7664The Analog Devices AD7664 ADC was tested for SEU and
SEL at BNL using ions with LETs ranging from 7.88 (Si) to37.9 (Br). The ADCs were tested under the bias conditions of+ 5 V.
Testing was performed to determine the latchup thresholdLET. Both devices tested showed the same thresholdcharacteristics with an LET threshold of 8-10 MeV•cm2/mg.For all test conditions, the DUTs were exposed to the ionbeam until latchup occurred. At this point the fluence wasrecorded. Typically, 2 to 3 fluence readings were taken foreach test condition to allow for some statistics in calculating across section. This cross section data is shown in Figure 18.
Consider the cross section data in Figure 18. SEL did notoccur for 5 x 106 ions/cm2 of Si at normal incidence(LET=7.88). The data point with the downward arrows forthis LET value thus represent upper limits for the SEL crosssection. SEL did occur for Si incident at 30o (LET=9.1),indicating an SEL threshold LET between these two values.Despite the considerable spread in the cross section data, the“best fit” curve through the data yields a reasonable estimateof the saturation cross section, σsat ≅ 2 to 3 x 10-4 cm2.
It should be noted that for LETth of 10 or less, thepossibility of sensitivity to proton-induced events exists. Thispossibility is not addressed by this testing. [28]
10-7
10-5
10-4
10-3
4030
DUT #1 DUT #2 Best Exponential Fit
Latc
hup
Cro
ss S
ectio
n (c
m2)
Effective LET (MeV-cm2/mg)
10 200
10-6
Figure 18. The symbols show the cross section data as a function ofEffective LET for the two DUTs. Note: Symbols with a downward arrowindicate that no SEL occurred for that LET.
4) AD7854The Analog Devices AD7854 is a CMOS single-channel,
200 kbps 12 bit ADC. It was tested for SEL only and foundto be susceptible to SEL with a threshold LET between 8 and11 MeV•cm2/mg and with a limiting cross section between1x10-4 and 1x10-3 cm2. The device was susceptible for both3.6 and 5 volt supply voltages. [29]
5) AD7858The Analog Devices AD7858 is a CMOS 8 channel, 200
kbps 12 bit ADC. It was tested for SEL only and found to besusceptible with a threshold LET between 11 and 22MeV•cm2/mg and a limiting cross section between 1x10-4 and1x10-3 cm2. The device was susceptible for both 3.6 and 5volt supply voltages. [30]
6) AD7888The Analog Devices AD7888 is a CMOS 8 channel, 125
kbps 12 bit ADC. It was tested for SEL only and found to besusceptible with a threshold LET between 16 and 22 and alimiting cross section between 1x10-5 and 1x10-4 cm2. Thedevice was susceptible for both 3.6 and 5 volt supplyvoltages. [31]
C. Miscellaneous:
1) ADG704The Analog Devices ADG704 is a low-voltage (1.8 to 5 V)
quad CMOS multiplexer. It was tested for SEL only andfound to be susceptible with a threshold LET of about 16 andwith a limiting cross section between 1x10-5 and 1x10-4 cm2.The device was tested only at the upper limit of its supplyvoltage range. [32]
2) AMP01The AMP01 was tested to further characterize a destructive
failure observed in 1998 SEE tests [5]. The new tests raisedthe lower limit on the LET threshold for failure to >32.7(1x107 ions/cm2 of Ge at normal incidence). No dependenceon temperature or device load were observed. The higherthreshold LET and the fact that failure is more likely fornormally incident ions imply that the failure rate for thismechanism will be low – for example, less than 1x10-10
failures per day per device in geostationary orbit. [5].
3) INA117The Texas Instruments/Burr-Brown INA117 is a
differential amplifier. It was tested for SEL and otherdestructive mechanisms only. No destructive events wereobserved for an effective fluence of 1x107 ions/cm2 of Iodineincident at 60 degrees for an effective LET of 119.9.
4) MAX313Testing was performed at BNL to determine the SEL
susceptibility of the Maxim MAX313 Analog Switch. Supplycurrent was monitored for an increase or decrease. No SELswere observed for the MAX313 up to a fluence of1x107 ions/cm2 of iodine at normal incidence (LET=59.8).[33]
5) MAX4503The Maxim MAX4503 CMOS switch was tested for SEL
susceptibility at BNL. Test conditions included the supplyvoltage set to 3.6 V, the DC input to the switch of 2.15 V.The switching voltage set to 3.3 V and the frequency was 200Hz. No SELs were observed up to LET of 37. At LET valuesof 37 and greater, transients on the output were beginning to
appear. As this was only a latchup test, details of thetransients were not investigated. However, with an LETthreshold of 37, the event rate should be very small. [34]
6) MAX4528 and MAX4583SEL sensitivity tests were preformed at BNL on the Maxim
MAX4528 and MAX 4583 CMOS analog switches. Thepower supply current was monitored for large increases andthe device functionality was verified after each SEL.
The MAX4528 exhibited no SEL events up to a fluence of1x107 ions/cm2 of iodine (LET=59.9). The MAX4583exhibited no SEL events for a fluence of 1x107 ions/cm2 ofiodine incident at 60o (LETeff=119.8). [35]
7) MAX4617The Maxim MAX4617 CMOS analog multiplexer was
tested at BNL for sensitivity of SEL. The power supply wasmonitored for large increases. The MAX4617 exhibited noSEL events for a fluence of 1x107 ions/cm2 of iodine incidentat 60o (LETeff=119.8). [36]
8) MAX584Testing was performed at BNL to determine the SEL
susceptibility of the Maxim MAX584 voltage reference. NoSELs were observed up to a fluence of 1x107 ions/cm2 of Brat normal incidence (LETeff=37.3).
VII. SUMMARY
We have presented recent data from SEE, and proton-induced damage tests on a variety of mainly commercialdevices. It is the authors’ recommendation that this data beused with caution. We also highly recommend that lot testingbe performed on any suspect or commercial device.
VIII. ACKNOWLEDGMENT
The Authors would like to acknowledge the sponsors ofthis effort: NASA Electronics Radiation Characterization(ERC) Project, a portion of NASA Electronic Parts andPackaging Program (NEPP), NASA Flight Projects, and theDefense Threat Reduction Agency (DTRA) under IACRO01-4050/0001278.
IX. REFERENCES
[1] NASA/GSFC Radiation Effects and Analysis home page,http://radhome.gsfc.nasa.gov
[2] K. LaBel, private communication, May 2001.[3] C Poivey, et al., “Development of a Test Methodology for
Single Event Transients (SETs) in Linear Devices,” submittedand accepted for publication in 2001 IEEE Radiation EffectsData Workshop, July, 2001, PH-2.
[4] J.W. Howard, et al., “Total Dose and Single Event EffectsTesting of the Intel Pentium III (P3) and AMD K7Microprocessors,” submitted and accepted for publication in2001 IEEE Radiation Effects Data Workshop, July, 2001, W-7.
[5] M.V. O'Bryan, et al., “Recent Radiation Damage and SingleEvent Effect Results for Microelectronics,” NSREC99 DataWorkshop, 1-14, July 1999.
[6] Jim Howard, et al., "Single Event Upset and Latchup Testing ofthe AD1674 Analog Devices Analog to Digital Converter,"
[7] Jim Howard, et al., "Single Event Upset and Latchup Testing ofthe AD6640 Analog Devices Analog to Digital Converter,"http://radhome.gsfc.nasa.gov/radhome/papers/B111600a.pdf,November 2000.
[8] R. A. Reed, et al., "Proton-Induced Single Event UpsetCharacterization of a 1 Giga-Sample per Second Analog toDigital Converter," Fifth European Conf. Radiation Effects onComponents and Systems, RADECS 99, 188-192, September1999.
[9] Jim Howard, et al., "Single Event Transient and DestructiveSingle Event Effects Testing of the AFL12005SX/CHAdvanced Analog Lambda, Inc. DC/DC Converters,"http://radhome.gsfc.nasa.gov/radhome/papers/B080100a.pdf,August 2000.
[10] Jim Howard, et al., "Single Event Transient and DestructiveSingle Event Effects Testing of the AFL12012DX/CHAdvanced Analog Lambda, Inc. DC/DC Converters,"http://radhome.gsfc.nasa.gov/radhome/papers/B080100b.pdf,August 2000.
[11] Jim Howard, et al., "Single Event Transient and DestructiveSingle Event Effects Testing of the AFL12015DX/CHAdvanced Analog Lambda, Inc. DC/DC Converters,"http://radhome.gsfc.nasa.gov/radhome/papers/B080100c.pdf,August 2000.
[12] Jim Howard, et al., "Single Event Transient and DestructiveSingle Event Effects Testing of the DVHF2803R3SF/HBMVPT, Inc. DC/DC Converters," http://radhome.gsfc.nasa.gov/radhome/papers/B111500a.pdf, November 2000.
[13] Jim Howard, et al., "Single Event Transient and LatchupTesting of the LCM-120 Interpoint Line ConditioningModule," http://radhome.gsfc.nasa.gov/radhome/papers/B080100d.pdf, August 2000.
[14] A. Sanders, "Total Ionizing Dose (TID) Test Report, Version1.0 Interpoint Line Conditioning Module (LCM) LCM-120SN0025 T0003 LCM-120 SN0011 T9952 Tested 5/22/00-5/27/00," http://radhome.gsfc.nasa.gov/radhome/papers/G052200.pdf, July 2000.
[15] Jim Howard, et al., "Single Event Transient and DestructiveSingle Event Effects Testing of the MDI3051RES05ZFModular Devices, Inc. DC/DC Converters",http://radhome.gsfc.nasa.gov/radhome/papers/B080100e.pdf,August 2000.
[16] Jim Howard, et al., "Single Event Transient and DestructiveSingle Event Effects Testing of the MDI3051RES12ZFModular Devices, Inc. DC/DC Converters,"http://radhome.gsfc.nasa.gov/radhome/papers/B080100f.pdf,August 2000.
[17] Jim Howard, et al., "Single Event Transient and DestructiveSingle Event Effects Testing of the MDI3051RES15ZFModular Devices, Inc. DC/DC Converters,"http://radhome.gsfc.nasa.gov/radhome/papers/B080100g.pdf,August 2000.
[18] S. Kniffin et al., "Heavy Ion and Proton Single Event EffectsTest Results for the Fairchild LM148 Quad Op Amp,"http://radhome.gsfc.nasa.gov/radhome/papers/D082400.pdf,October 2000.
[19] Jim Howard, et al., "Proton Dose and Single Event EffectsTesting of the Intel Pentium III (P3) and AMD K7Microprocessors," http://radhome.gsfc.nasa.gov/radhome/papers/i062100.pdf, June 2000.
[20] Robert Reed and Hak Kim, "Heavy Ion Single Event EffectsTest Results for the NonVolatile Electronics IL710 Isolator",
[21] S. Kniffin et al., "Heavy Ion Single Event Effects Test Resultsfor the Micropac Mii42142 Power Operational Amplifier,"http://radhome.gsfc.nasa.gov/radhome/papers/D120400c.pdf,September, 2000.
[22] S. Kniffin et al., "Heavy-Ion and Proton Test Results for theMicropac Mii53124 and Mii53253 Power MOSFETOptocouplers," http://radhome.gsfc.nasa.gov/radhome/papers/D120400e.pdf, September, 2000.
[23] S. Kniffin et al., "Heavy Ion Single Event Effects Test Resultsfor the Micropac Mii53250 and Mii53258 Optocoupler-BasedSolid State Relays," http://radhome.gsfc.nasa.gov/radhome/papers/D120400d.pdf, September, 2000.
[24] S. Kniffin et al., "Test Report for STRV-1d Ground Data Takenat UC Davis," http://radhome.gsfc.nasa.gov/radhome/papers/D120400a.pdf, December 2000.
[25] R.A. Reed, C.J. Marshall, J.L. Barth, K.A. LaBel, C. Poivey,P.W. Marshall, S. Kniffin, C. Seidleck, “Assessing the Impactof the Space Radiation Environment on ParametricDegradation and Single Event Transients in Optocouplers,”submitted and accepted for publication in 2000 IEEE RadiationEffects Data Workshop, July, 2001, PH-5.
[26] S. Kniffin et al., "Proton Effects Test Results for the OptodiodeOD800 Light Emitting Diode," http://radhome.gsfc.nasa.gov/radhome/papers/D080700.pdf, August 2000.
[27] S. Kniffin et al., "Report for VCE Degradation in Mii 4N49Optocouplers for IRAC," http://radhome.gsfc.nasa.gov/radhome/papers/D120400b, December 2000.
[28] Jim Howard and Jim Fourney, "Single Event Latchup Testingof the AD7664 Analog Devices Analog to Digital Converter,"http://radhome.gsfc.nasa.gov/radhome/papers/B030600.pdf,March 2000.
[29] Ray Ladbury, et al., "Heavy Ion Latch-up Test Results for theAnalog Devices AD7854 ADC," http://radhome.gsfc.nasa.gov/radhome/papers/B092500a.pdf, September 2000.
[30] Ray Ladbury, et al., "Heavy Ion Latch-up Test Results for theAnalog Devices AD7858 ADC," http://radhome.gsfc.nasa.gov/radhome/papers/B092500b.pdf, September 2000.
[31] Ray Ladbury, et al., "Heavy Ion Latch-up Test Results for theAnalog Devices AD7888 ADC," http://radhome.gsfc.nasa.gov/radhome/papers/B092500c.pdf, September 2000
[32] Ray Ladbury, et al., "Heavy Ion Latch-up Test Results for theAnalog Devices ADG704 Multiplexer," http://radhome.gsfc.nasa.gov/radhome/papers/B092700b.pdf, September 2000.
[33] Jim Howard, et al., "Heavy Ion Single Event Transient (SET)and Latchup (SEL) Test Results for the Maxim Analog Switch(MAX313MTE)," http://radhome.gsfc.nasa.gov/radhome/papers/B080100h.pdf, August 2000.
[34] Jim Howard, et al., "Heavy Ion Single Event Latchup (SEL)Test Results for the Maxim CMOS Switch (MAX4503),"http://radhome.gsfc.nasa.gov/radhome/papers/B111600c.pdf,November 2000.
[35] Ray Ladbury, et al., "Heavy Ion Latch-up Test Results for theMaxim MAX4528, MAX4583 CMOS Analog Switches,"http://radhome.gsfc.nasa.gov/radhome/papers/B092700c.pdf,September 2000.
[36] Ray Ladbury, et al., "Heavy Ion Latch-up Test Results for theMaxim MAX4617 CMOS Analog Multiplexer,"http://radhome.gsfc.nasa.gov/radhome/papers/B111700.pdf,November 2000.
![Spacecraft Simulation]](https://static.documents.pub/doc/80x56/544e0a73b1af9f33638b4bf0/spacecraft-simulation.jpg)
